Metformin prescription patterns among US adolescents aged 10-19 years: 2009-2013.

WHAT IS KNOWN AND OBJECTIVE: Metformin is the only oral antihyperglycemic agent approved for use in adolescents with type 2 diabetes mellitus (T2DM). There are reports of metformin used to treat conditions such as obesity, hyperinsulinemia, prediabetes, metabolic syndrome and polycystic ovarian syndrome (PCOS). It is important to understand metformin prescription patterns and underlying diagnoses in adolescents as it can provide estimates of the extent of on-label (i.e. treatment of T2DM) and off-label use of metformin in this population. Our study sought to assess metformin prescription patterns among US adolescents from 2009 to 2013. METHODS: Data from the National Disease and Therapeutic Index (NDTI) database, the MarketScan(®) Commercial Claims and Encounters database and the Multi-State Medicaid database were analysed. The proportion of diagnoses associated with metformin that was recommended during a clinical visit was identified in the NDTI database. In the MarketScan(®) Commercial and Medicaid databases, adolescents with at least one metformin prescription with ±6 months continuous enrolment from the date of the index metformin prescription were included in the analyses. All diagnosis and procedure codes were extracted within ±6 months of the index metformin prescription. The proportion of T2DM was calculated irrespective of any other medical conditions, whereas all other prespecified conditions were classified as positive only if no concurrent T2DM diagnosis codes were present. RESULTS AND DISCUSSION: In the NDTI database, the most common diagnoses associated with metformin use were diabetes (34·9%), followed by metabolic syndrome (20·9%), PCOS (17·2%) and obesity (6·5%). In the MarketScan(®) Commercial database, T2DM was the most common diagnosis among girls aged 10-14 years (22·8-23·6%), boys aged 10-14 years (20·5-24·5%) and boys aged 15-19 years (37·1-43·1%), whereas PCOS (24·1-28·3%) was the most common diagnosis among girls aged 15-19 years. In the Medicaid database, T2DM was the most common diagnosis among all four groups and the proportions were higher than their counterparts in the Commercial database. WHAT IS NEW AND CONCLUSION: Analyses from three separate US data sources suggest that off-label prescribing of metformin is common among US adolescents aged 10-19 years. To avoid potential overestimation, caution should be exercised when utilizing metformin prescription as a proxy measure to estimate the burden of T2DM in adolescents.

Recent trends in the characteristics of patients prescribed sitagliptin and other oral antihyperglycaemic agents in a large U.S. claims database.

AIM: This study was designed to determine if differences in baseline characteristics of patients with type 2 diabetes mellitus (T2DM) being treated with sitagliptin vs. other oral antihyperglycaemic agents (OAHA) during the initial 2 years following sitagliptin's introduction in the U.S. continued during the second 2 years of sitagliptin availability. METHODS: Patients with T2DM and at least one new prescription for sitagliptin or another OAHA from Oct 2006 to April 2010 were identified in an insurance claims database. Multivariate logistic regression adjusting for age, gender, treatment type (monotherapy, dual or triple therapy), new or existing T2DM diagnosis, and comorbidities and diabetes complications in the prior 12 months was used to estimate odds ratios for sitagliptin vs. other OAHAs. RESULTS: During 2006-2007 or 2008-2010, new sitagliptin users were older and more likely to be male, have prior diagnosis of T2DM, or initiating combination therapy compared with new users of other OAHAs. Prevalence of comorbidities and complications was consistently higher for new sitagliptin users across most of the conditions assessed during both time periods. CONCLUSIONS: New sitagliptin users consistently tended to be older and have greater comorbidity/complication burden compared with new users of other OAHAs. These differences in baseline characteristics persisted up to 4 years postapproval. This observation has significant implications for observational studies using electronic medical record or insurance claims databases. Appropriate adjustment is needed to try to control for potential confounding and channelling bias resulting from this non-random prescribing pattern, and the limitations of such analyses acknowledged.